A processing unit for a multiprocessor data processing system includes a store-through upper level cache, an instruction sequencing unit that fetches instructions for execution, at least one instruction execution unit that executes a store-conditional instruction to determine a store target address, a store queue that, following execution of the store-conditional instruction, buffers a corresponding store operation, sequencer logic associated with the store queue. The sequencer logic, responsive to receipt of a latency indication indicating that resolution of the store-conditional operation as passing or failing is subject to significant latency, invalidates, prior to resolution of the store-conditional operation, a cache line in the store-through upper level cache to which a load-reserve operation previously bound.
|
12. A method of processing in a multiprocessor data processing system including a memory hierarchy including a store-through upper level cache, a lower level cache and a system memory, said method comprising:
executing a store-conditional instruction to determine a store target address;
following execution of said store-conditional instruction, buffering a corresponding store operation; and
responsive to receipt of a latency indication indicating that resolution of said store-conditional operation as passing or failing is subject to significant latency, a processing core invalidating, prior to resolution of said store-conditional operation, a cache line in the store-through upper level cache to which a load-reserve operation previously bound.
1. A processing unit for a multiprocessor data processing system, said processing unit comprising:
a store-through upper level cache;
an instruction sequencing unit that fetches instructions for execution;
at least one instruction execution unit that executes a store-conditional instruction to determine a store target address;
a store queue that, following execution of said store-conditional instruction, buffers a corresponding store operation; and
sequencer logic associated with the store queue that, responsive to receipt of a latency indication indicating that resolution of said store-conditional operation as passing or failing is subject to significant latency, invalidates, prior to resolution of said store-conditional operation, a cache line in said store-through upper level cache to which a load-reserve operation previously bound.
7. A data processing system, comprising:
a system memory;
a system interconnect; and
a plurality of processing units coupled to the system interconnect, at least a particular processing unit among said plurality of processing units including:
a store-through upper level cache;
an instruction sequencing unit that fetches instructions for execution;
at least one instruction execution unit that executes a store-conditional instruction to determine a store target address;
a store queue that, following execution of said store-conditional instruction, buffers a corresponding store operation; and
sequencer logic associated with the store queue that, responsive to receipt of a latency indication indicating that resolution of said store-conditional operation as passing or failing is subject to significant latency, invalidates, prior to resolution of said store-conditional operation, a cache line in said store-through upper level cache to which a load-reserve operation previously bound.
2. The processing unit of
4. The processing unit of
5. The processing unit of
6. The processing unit of
said at least one instruction execution unit executes the load-reserve instruction to determine a load target address; and
the processing unit, responsive to the execution of the load-reserve instruction, performs a corresponding load-reserve operation by accessing said store-through upper level cache utilizing said load target address to cause data associated with said load target address to be loaded from said store-through upper level cache into a data register and by establishing a reservation for a reservation granule including the load target address.
8. The data processing system of
9. The data processing system of
10. The data processing system of
11. The data processing system of
said at least one instruction execution unit executes the load-reserve instruction to determine a load target address; and
the particular processing unit, responsive to the execution of the load-reserve instruction, performs a corresponding load-reserve operation by accessing said store-through upper level cache utilizing said load target address to cause data associated with said load target address to be loaded from said store-through upper level cache into a data register and by establishing a reservation for a reservation granule including the load target address.
13. The method of
14. The method of
15. The method of
16. The method of
executing the load-reserve instruction to determine a load target address; and
the processor core, responsive to the execution of the load-reserve instruction, performing a corresponding load-reserve operation by accessing said store-through upper level cache utilizing said load target address to cause data associated with said load target address to be loaded from said store-through upper level cache into a data register and by establishing a reservation for a reservation granule including the load target address.
|
The present application is related to the following applications, which are filed on even date herewith, assigned to the same assignee, and incorporated herein by reference in their entireties:
1. Technical Field
The present invention relates in general to data processing system and, in particular, to managing accesses to shared data in a data processing system. Still more particularly, the present invention relates to a processor, data processing system and method for synchronizing accesses to data in shared memory.
2. Description of the Related Art
In shared memory multiprocessor (MP) data processing systems, each of the multiple processors in the system may access and modify data stored in the shared memory. In order to synchronize access to a particular granule (e.g., cache line) of memory between multiple processing units and threads of execution, load-reserve and store-conditional instruction pairs are often employed. For example, load-reserve and store-conditional instructions have been implemented in the PowerPC® instruction set architecture with operation codes (opcodes) associated with the LWARX and STWCX mnemonics, respectively (referred to hereafter as LARX and STCX). The goal of load-reserve and store-conditional instruction pairs is to load and modify data and then to commit the modified data to coherent memory only if no other thread of execution has modified the data in the interval between the load-reserve and store-conditional instructions. Thus, updates to shared memory can be synchronized without the use of an atomic update primitive that strictly enforces atomicity.
With reference now to
As further illustrated in
L1 cache 126 is a store-through cache, meaning that the point of cache coherency with respect to other processing units 102 is below the L1 cache (e.g., at L2 cache 130). L1 cache 126 therefore does not maintain true cache coherency states (e.g., Modified, Exclusive, Shared, Invalid) for its cache lines, but only maintains valid/invalid bits. Store operations first complete relative to the associated processor core 120 in the L1 cache and then complete relative to other processing units 102 in L2 cache 130.
As depicted, in addition to the L2 cache array 140, L2 cache 130 includes read-claim (RC) logic 142 for managing memory access requests by the associated processor core 120, snoop logic 144 for managing memory access requests by other processing units 102, and reservation logic 146 for recording reservations of the associated processor core 120. Reservation logic 146 includes at least one reservation register comprising a reservation address field 148 and a reservation flag 150.
Following block 202, the process passes to block 204, which illustrates processor core 120 issuing a LARX operation corresponding to the LARX instruction to RC logic 142 within L2 cache 130. As depicted at bock 206, RC logic 142 stores the address of the reservation granule (e.g., cache line) containing the target address in reservation address field 148 and sets reservation flag 150. Reservation logic 146 then begins monitoring for an indication by snoop logic 144 that another processing unit 102 has updated the cache line containing the target address. As illustrated at block 208, RC logic 142 obtains the load data from L2 cache array 140, system memory 108 or another processing unit 102 and then returns the requested load data to processor core 120. In response to receipt of the load data, processor core 120 stores the load data in an internal register, but not in L1 cache 126.
Processor core 120 thereafter attempts to perform an atomic update to the load data through the execution of a store-conditional (e.g., STCX) instruction in accordance with the process depicted in
Following block 222, processor core 120 issues a store-conditional (e.g., STCX) operation corresponding to the store-conditional instruction to RC logic 142 within L2 cache 130, as shown at block 224. RC logic 142 obtains owner permission for the target cache line and then determines at block 226 whether or not reservation flag 150 is still set (i.e., whether or not any other processing unit 102 has modified the reservation granule). If reservation flag 150 is still set, indicating that no other processing unit 102 has modified the reservation granule, RC logic 142 updates L2 cache array 140 with the store data and resets reservation flag 150, as shown at block 228. Reservation logic 146 then sends a pass indication to processor core 120, as shown at block 230. Thereafter, the process ends at block 234.
Returning to block 226, in response to a determination that reservation flag 150 is reset, indicating that another processing unit 102 has modified the reservation granule in the interval between execution of the load-reserve and store-conditional instructions, the store-conditional operation fails in L2 cache 130, and reservation logic 146 transmits a fail indication to processor core 120, as depicted at block 232. Thereafter, processing of the store-conditional operation terminates at block 234.
The present invention appreciates that the above-described conventional technique of updating shared memory utilizing load-reserve and store-conditional instructions has a number of drawbacks. For example, the conventional methodology described with respect to
In addition, the conventional technique of updating shared memory utilizing load-reserve and store-conditional instructions disadvantageously results in the invalidation of the target cache line in L1 cache 126. As will be appreciated, one advantage of caching data is that software often exhibits locality of reference, meaning that data within the same cache line as the data updated by the load-reserve/store-conditional pair is frequently referenced by other nearby instructions. Unfortunately, the invalidation of the cache line in L1 cache 126 delays the execution of such instructions until the cache line is loaded back into L1 cache 126 from L2 cache 130.
In view of these and additional shortcomings in the prior art, the present invention provides improved processors, data processing systems and methods for synchronizing updates to a shared memory utilizing load-reserve and store-conditional instructions. According to one embodiment, A processing unit for a multiprocessor data processing system includes a store-through upper level cache, an instruction sequencing unit that fetches instructions for execution, at least one instruction execution unit that executes a store-conditional instruction to determine a store target address, a store queue that, following execution of the store-conditional instruction, buffers a corresponding store operation, sequencer logic associated with the store queue. The sequencer logic, responsive to receipt of a latency indication indicating that resolution of the store-conditional operation as passing or failing is subject to significant latency, invalidates, prior to resolution of the store-conditional operation, a cache line in the store-through upper level cache to which a load-reserve operation previously bound.
All objects, features, and advantages of the present invention will become apparent in the following detailed written description.
The novel features believed characteristic of the invention are set forth in the appended claims. However, the invention, as well as a preferred mode of use, will best be understood by reference to the following detailed description of an illustrative embodiment when read in conjunction with the accompanying drawings, wherein:
With reference again to the figures and in particular with reference to
As further illustrated in
Each processor core 320 further includes an L1 store queue (STQ) 327 and a load unit 328 for managing the completion of store and load operations, respectively, corresponding to executed store and load instructions (including load-reserve and store-conditional instructions). In a preferred embodiment, L1 STQ 327 is implemented as a First-In, First-Out (FIFO) queue containing a plurality of queue entries. Store operations are accordingly loaded in the “top” entry of L1 STQ 327 at execution of the corresponding store instruction to determine the target address, and are initiated when the store operation reaches the “bottom” or “commit” entry of L1 STQ 327.
It is important to note that the present application makes a distinction between “instructions”, such as load-reserve and store-conditional instructions, and “operations”. Load and store “instructions” (including load-reserve and store-conditional instructions) are defined herein as inputs to an execution unit that include an operation code (opcode) identifying the type of instruction and one or more operands specifying data to be accessed and/or its address. Load and store “operations”, including load-reserve and store-conditional operations, are defined herein as data and/or signals generated following instruction execution that specify the address of data to be accessed. Thus, load-reserve and store-conditional operations may be transmitted from a processor core 320 to lower level memory to initiate data accesses, while load-reserve and store-conditional instructions are not.
The operation of processor core 320 is supported by a multi-level volatile memory hierarchy having at its lowest level shared system memory 308, and at its upper levels two or more levels of cache memory, which in the illustrative embodiment include L1 cache 326 and L2 cache 330. In the particular embodiment shown in
In accordance with the present invention, L1 cache 326, which may include bifurcated L1 data and instruction caches, is implemented as a store-through cache, meaning that the point of cache coherency with respect to other processor cores 320 is located below L1 cache 326 and, in the depicted embodiment, is located at L2 cache 330. Accordingly, as described above, L1 cache 326 does not maintain true cache coherency states (e.g., Modified, Exclusive, Shared, Invalid) for its cache lines, but only maintains valid/invalid bits. Because L1 cache 326 is implemented as a store-through cache, store operations (including store-conditional operations) first complete relative to the associated processor core 120 in L1 cache 326 and then complete relative to other processing units 302 at a point of system-wide coherency, which in the depicted embodiment is L2 cache 330.
As further illustrated in
L2 cache 330 further includes a number of instances of snoop (SN) logic 344a-344n for managing memory access and other operations received from other processing units 302 via system interconnect 304 and snoop bus 370. Snoop logic instances 344 and RC logic instances 342 are each connected to a back-invalidation bus 372 by which any snoop logic instance 344 or RC logic instance 342 can signal the invalidation of a cache line to processor core 320.
L2 cache 330 finally includes reservation logic 346 for recording reservations of the associated processor core(s) 320. Specifically, reservation logic 346 includes, for each thread that may be concurrently executed by the processor core(s) 320 in its processing unit 302, a respective reservation register comprising a reservation address field 348 and a reservation flag 350. For example, in the depicted example, which assumes that processor cores 320a and 320b can each execute a single thread, reservation logic 346 includes two reservation registers: reservation address field 348a and reservation flag 350a for processor core 320a and reservation address field 348b and reservation flag 350b for processor core 320b. When set (e.g., to ‘1’), a reservation flag 350 indicates that the associated processor core 320 holds a reservation for the address contained in reservation address field 348 and otherwise indicates no reservation is held. Reservation logic 346 supplies pass/fail indications indicating the outcomes of store-conditional operations to processor cores 320a, 320b via respective pass/fail buses 374a, 374b.
Referring now to
Following execution of the LARX instruction, the load target address is received from execution units 324 by LD unit 328. As illustrated at block 404, LD unit 328 then accesses the directory of L1 cache 326 to determine whether a cache line containing the data word(s) associated with the load target address resides in L1 cache 326. If not, L1 cache 326 signals a cache miss to LD unit 328, which responds by issuing a LARX operation to an RC logic instance 342 via load bus 360 and command bus 362, as shown at block 410.
Next, at block 412, the RC logic instance 342 initiates an access to L2 storage array and directory 340 to retrieve the cache line of data identified by the load target address and returns the requested cache line to processor core 320. Processor core 320 loads the entire cache line of data within L1 cache 326 and transfers the data word(s) associated with the load target address in a core register 323. Thus, unlike prior art processors that only load the data word(s) associated with the load target address in core register(s) and do not cache the target cache line, processor core 320 of data processing system 300 establishes the target cache line in L1 cache 326 to permit subsequent loads, and particularly subsequent load-reserve operations, to hit in L1 cache 326. Next, as illustrated at block 414, the RC logic instance 342 establishes a reservation for the load target address in L2 cache 330 by placing the load target address in a reservation address field 348 and setting the associated reservation flag 350. Thereafter, reservation logic 346 begins monitoring for an indication that the reservation granule has been updated, and the process terminates at block 416.
Returning to block 404, in response to LD unit 328 determining that the load target address hit in L1 cache 326, LD unit 328 issues an address-only LARXA operation to an RC logic instance 342 via load bus 360 and command bus 362 to inform L2 cache 330 that the load-reserve has bound to a (potentially stale) value in L1 cache 326. If processing unit 302 implements multiple processor cores 320 and/or processor core(s) 320 capable of executing multiple concurrent threads of execution, the LARXA operation preferably includes not only the load target address, but also the core ID and thread ID that executed the LARX instruction.
As shown at block 408, L1 cache 326 loads the data word(s) associated with the load target address into one or more core registers 323 associated with execution units 324. Thereafter, the process passes to block 414, which has been described.
In order to ensure that the reservation established at block 414 is properly subject to any snoop operations that may cancel the reservation, the interval between processing of the LARXA operation by LD unit 328 at block 404 to the presentation of the LARXA operation on load bus 360 as depicted at block 406 must be a short, bounded time. It is not required, however, that this interval be fixed to a particular number of cycles. To ensure the time to presentation of a LARXA operation is bounded, execution units 324 ensure that no other load (including load-reserve) instructions are executed in any thread in the same processor core 320 within one L2 (not core) cycle of the execution of the load-reserve instruction. This condition guarantees that load bus 360 will be available for LD unit 328 to transmit the LARXA operation to L2 cache 330 in a non-blocking fashion. In addition, LD unit 328 ignores any busy signals on load bus 360 for purposes of presenting LARXA operations because LARXA operations are processed only by reservation logic 346 (and not RC logic instances 342) and therefore do not need to be delayed, even if RC logic instances 342 are all busy.
With reference now to
Following block 422, the process proceeds to block 424, which depicts L1 STQ 327 issuing a STCX operation to L2 STQ 366 via store bus 364 when the STCX operation reaches the bottom (or “commit”) entry of L1 STQ 327. In addition, as depicted at block 426, L1 STQ 327 determines, when the STCX operation reaches the commit entry of L1 STQ 327, whether the STCX operation hit in L1 cache 326. If the store target address misses in L1 cache 326 at the commit entry of L1 STQ 327, the process passes to block 430, which depicts L1 STQ 327 awaiting a pass/fail indication indicating whether or not the STCX operation succeeded at L2 cache 330 (i.e., with respect to all other threads and cache coherent processing units 302). In response to receipt of the pass/fail operation, processor core 320 provides the pass/fail indication to execution units 324 (e.g., to indicate whether the path of execution should change), as shown at block 432, but does not write to L1 cache 326. Thereafter, the STCX operation is deallocated from L1 STQ 327, and the process terminates at block 434.
Returning to block 426, in response to L1 STQ 327 determining that the store target address of the STCX operation hit in L1 cache 326 at the commit entry, the process proceeds to block 440. Block 440 illustrates L1 STQ 327 initiating sequencer logic 325 to manage the conditional update to L1 cache 326. That is, because the success of the STCX operation is conditioned on the absence of other store operations to the same reservation granule following the point in time that the associated LARX operation binds to a value, the store data of the STCX operation cannot be placed in the L1 cache until the L2 has processed the STCX operation and ensured that it has completed successfully system-wide.
As noted above, L1 STQ 327 is preferably implemented as a FIFO queue, meaning that until the STCX operation in the bottom or “commit” entry of L1 STQ 327 has finished and written to L1 cache 326 (if necessary), L1 STQ 327 is blocked and other store operations cannot make forward progress. In order to avoid a potential deadlock and/or performance degradation, sequencer logic 325 of L1 STQ 327 preferably implements at least one mechanism by which sequencer logic 325 can discontinue waiting for the pass/fail indication from L2 cache 330 in order to update L1 cache 326 and can instead invalidate the target cache line in L1 cache 326.
In a preferred embodiment, sequencer logic 325 of L1 STQ 327 implements two “escape” mechanisms to cause sequencer logic 325 to stop waiting for the pass/fail indication from L2 cache 330 in order to update L1 cache 326. The first of these mechanism is a respective cancel (or latency) signal from L2 cache 330 for each processor core 320 and thread that, when asserted, indicates that L2 cache 330 is likely to delay returning the pass/fail indication (e.g., because of an L2 cache miss, the target cache line is held in Shared state, etc.). As shown at blocks 442 and 444, when sequencer logic 325 of L1 STQ 327 receives the cancel (latency) signal, sequencer logic 325 of L1 STQ 327 invalidates the target cache line in L1 cache 326 if the target cache line is still valid.
The other mechanism by which sequencer logic 325 can abandon the update to L1 cache 326 is an optional timeout counter (T) 329. If implemented, L1 STQ 327 starts timeout counter 329 in response to issuance of a STCX operation to L2 cache 330, and in response to a predetermined, programmable number of cycles elapsing prior to receipt of a pass/fail indication from L2 cache 330 (as indicated by a latency signal generated by timeout counter 329), sequencer logic 325 similarly invalidates the target cache line in L1 cache 326, as shown at block 442 and 444. By abandoning the writeback of the store data to L1 cache 326 in this manner, L1 STQ 327 can continue to make forward progress on other store operations. The process thereafter proceeds from block 444 to block 430, which has been described.
Still referring to
Referring now to
Block 468 illustrates the RC logic instance 342 determining whether or not a reservation flag 350 for the store target address is set (i.e., whether or not any other processing core 320 or thread has modified the reservation granule). If not, reservation logic 346 transmits a fail indication to processor core 320 via pass/fail bus 374, as shown at block 470. Thereafter, the process terminates at block 476. If, on the other hand, a determination is made at block 468 that reservation flag 350 is still set, indicating that no other processor core 320 or thread has modified the reservation granule, RC logic instance 342 updates the L2 cache array within storage array and directory 340 with the store data of the STCX operation and resets reservation flag 350, as shown at block 472. Reservation logic 346 also sends a pass indication to processor core 320 at block 474. Thereafter, the process ends at block 476.
With reference now to
If, however, the snoop logic instance 342 determines at block 484 that the snooped operation is a storage-modifying operation, snoop logic instance 344 makes a further determination whether the target address of the storage-modifying operation matches the contents of a reservation address field 348. If not, the process passes to block 486, which has been described. If, however, the target address of the snooped storage-modifying operation matches an address contained in a reservation address field 348, snoop logic instance 344 resets the associated reservation flag 350 to cause any subsequent STCX operation to the address specified in reservation address field 348 to fail, as shown at block 490. In addition, the snoop logic instance 344 issues a back-invalidate signal specifying the target address of the snooped storage-modifying operation to processor core 320 (e.g., L1 cache 326) via back-invalidation bus 372, as shown at block 492. The back-invalidate signal clears the associated cache line from L1 cache 326, thus preventing subsequent LARX operations from hitting in the cache line. It should be noted that while the back-invalidation is underway, RC logic instances 342 and snoop logic instances 344 prevent any STCX or LARX (not LARXA) operations from proceeding until reservation flag 350 is reset and the cache line is invalidated in L1 cache 326.
Following issuance of the back-invalidate signal, the snoop logic instance 344 waits a fixed interval of time, as illustrated at block 494, to ensure that all (now stale) LARXA operations currently in flight to the given cache line have set a reservation flag 350 in reservation logic 346. This wait interval is accordingly equal to or greater than the bounded interval in which LD unit 328 issues LARXA operations to L2 cache 330 via load bus 360. In response to the wait interval elapsing, the snoop logic instance 344 resets the reservation flag 350, if any, associated with a reservation address field 348 containing an address matching the target address of the snooped storage-modifying operation. In a preferred embodiment, the signal that resets the reservation flag 350, if necessary, is a directory write operation that updates the L2 directory within storage array and directory 340 in response to the snooped operation. Thus, reservation flags 350 set by stale LARXA operations are reset in order to satisfy the architectural requirement that a STCX operation conditioned upon a stale LARX must fail. The process thereafter terminates at block 498.
It should be noted that in embodiments of the present invention in which multiple processor cores 320 share an L2 cache 330, as shown in
As has been described, the present invention provides improved processing units, data processing systems, and methods for synchronizing updates to shared memory utilizing load-reserve and store-conditional instructions. In accordance with the present invention, load-reserve operations are permitted to bind to possibly stale values in the L1 cache prior to determining whether the values are stale. If it is determined that the value to which the load-reserve operation bound in the L1 cache is stale, a subsequent store-conditional operation to the same reservation granule fails. If the value is determined to not be stale, the subsequent store-conditional operation succeeds, and the cache line containing the value to which the load-reserve operation bound is advantageously retained in the L1 cache.
While the invention has been particularly shown as described with reference to a preferred embodiment, it will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the spirit and scope of the invention. For example, although the invention has been described with reference to a particular embodiment of a memory hierarchy in which L1 and L2 caches are incorporated within a processing unit, those skilled in the art will appreciate that additional levels of in-line or lookaside caches may be employed and that, in some embodiments of the invention, one or more levels of cache may be off chip.
Williams, Derek Edward, Guthrie, Guy Lynn, Starke, William John, Levenstein, Sheldon B.
Patent | Priority | Assignee | Title |
10884740, | Nov 08 2018 | International Business Machines Corporation | Synchronized access to data in shared memory by resolving conflicting accesses by co-located hardware threads |
11068407, | Oct 26 2018 | International Business Machines Corporation | Synchronized access to data in shared memory by protecting the load target address of a load-reserve instruction |
11086672, | May 07 2019 | International Business Machines Corporation | Low latency management of processor core wait state |
11106608, | Jun 22 2020 | International Business Machines Corporation | Synchronizing access to shared memory by extending protection for a target address of a store-conditional request |
11119781, | Dec 11 2018 | International Business Machines Corporation | Synchronized access to data in shared memory by protecting the load target address of a fronting load |
11693776, | Jun 18 2021 | International Business Machines Corporation | Variable protection window extension for a target address of a store-conditional request |
8117397, | Dec 16 2008 | International Business Machines Corporation | Victim cache line selection |
8209489, | Oct 22 2008 | International Business Machines Corporation | Victim cache prefetching |
8225045, | Dec 16 2008 | International Business Machines Corporation | Lateral cache-to-cache cast-in |
8285939, | Apr 08 2009 | International Business Machines Corporation | Lateral castout target selection |
8296519, | Dec 31 2009 | International Business Machines Corporation | Synchronizing access to data in shared memory via upper level cache queuing |
8312220, | Apr 09 2009 | International Business Machines Corporation | Mode-based castout destination selection |
8327073, | Apr 09 2009 | International Business Machines Corporation | Empirically based dynamic control of acceptance of victim cache lateral castouts |
8347036, | Apr 09 2009 | International Business Machines Corporation | Empirically based dynamic control of transmission of victim cache lateral castouts |
8347037, | Oct 22 2008 | International Business Machines Corporation | Victim cache replacement |
8438569, | Oct 14 2004 | International Business Machines Corporation | Broadcasting a condition to threads executing on a plurality of on-chip processors |
8489819, | Dec 19 2008 | International Business Machines Corporation | Victim cache lateral castout targeting |
8499124, | Dec 16 2008 | International Business Machines Corporation | Handling castout cache lines in a victim cache |
8949540, | Mar 11 2009 | International Business Machines Corporation | Lateral castout (LCO) of victim cache line in data-invalid state |
9110808, | Dec 30 2009 | International Business Machines Corporation | Formation of an exclusive ownership coherence state in a lower level cache upon replacement from an upper level cache of a cache line in a private shared owner state |
9189403, | Dec 30 2009 | International Business Machines Corporation | Selective cache-to-cache lateral castouts |
9390026, | Feb 27 2014 | International Business Machines Corporation | Synchronizing access to data in shared memory |
9396127, | Feb 27 2014 | International Business Machines Corporation | Synchronizing access to data in shared memory |
9619390, | Dec 30 2009 | International Business Machines Corporation | Proactive prefetch throttling |
Patent | Priority | Assignee | Title |
20040073756, | |||
20050132148, |
Executed on | Assignor | Assignee | Conveyance | Frame | Reel | Doc |
Oct 07 2004 | GUTHRIE, GUY LYNN | International Business Machines Corporation | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 015300 | /0640 | |
Oct 07 2004 | WILLIAMS, DEREK EDWARD | International Business Machines Corporation | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 015300 | /0640 | |
Oct 08 2004 | STARKE, WILLIAM JOHN | International Business Machines Corporation | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 015300 | /0640 | |
Oct 12 2004 | LEVENSTEIN, SHELDON B | International Business Machines Corporation | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 015300 | /0640 | |
Oct 14 2004 | International Business Machines Corporation | (assignment on the face of the patent) | / | |||
May 03 2011 | International Business Machines Corporation | Google Inc | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 026664 | /0866 | |
Sep 29 2017 | Google Inc | GOOGLE LLC | CHANGE OF NAME SEE DOCUMENT FOR DETAILS | 044127 | /0735 |
Date | Maintenance Fee Events |
Jan 10 2011 | REM: Maintenance Fee Reminder Mailed. |
May 18 2011 | M1551: Payment of Maintenance Fee, 4th Year, Large Entity. |
May 18 2011 | M1554: Surcharge for Late Payment, Large Entity. |
Dec 05 2014 | M1552: Payment of Maintenance Fee, 8th Year, Large Entity. |
Jan 21 2019 | REM: Maintenance Fee Reminder Mailed. |
Jul 08 2019 | EXP: Patent Expired for Failure to Pay Maintenance Fees. |
Date | Maintenance Schedule |
Jun 05 2010 | 4 years fee payment window open |
Dec 05 2010 | 6 months grace period start (w surcharge) |
Jun 05 2011 | patent expiry (for year 4) |
Jun 05 2013 | 2 years to revive unintentionally abandoned end. (for year 4) |
Jun 05 2014 | 8 years fee payment window open |
Dec 05 2014 | 6 months grace period start (w surcharge) |
Jun 05 2015 | patent expiry (for year 8) |
Jun 05 2017 | 2 years to revive unintentionally abandoned end. (for year 8) |
Jun 05 2018 | 12 years fee payment window open |
Dec 05 2018 | 6 months grace period start (w surcharge) |
Jun 05 2019 | patent expiry (for year 12) |
Jun 05 2021 | 2 years to revive unintentionally abandoned end. (for year 12) |